SB 6120
In CommitteeSenate
High-risk AI
Regulating high-risk artificial intelligence system development, deployment, and use.
This status may be delayed. See Action History below for the latest updates.
How does a bill become law?
- Introduced: The bill is filed and assigned a number.
- Committee: A subject-matter committee holds hearings, takes public testimony, and decides whether to advance the bill.
- Floor Vote: The full chamber (House or Senate) debates and votes on the bill.
- Opposite Chamber: The bill repeats the committee and floor vote process in the other chamber.
- Governor: The Governor reviews the bill and decides whether to sign or veto it.
- Signed: The bill has been signed into law.
AI Analysis
This bill regulates the development and use of high-risk artificial intelligence systems in Washington to prevent algorithmic discrimination and ensure transparency. It requires developers and deployers to disclose how AI works, conduct risk assessments, and provide explanations for decisions that significantly affect people—like jobs, housing, or legal outcomes—while exempting certain sectors already under oversight.
- Creates a new legal category of 'high-risk artificial intelligence systems'—defined as AI used to make or significantly influence 'consequential decisions' (e.g., housing, employment, education, legal outcomes, insurance, health care).
- Requires developers to use reasonable care to prevent algorithmic discrimination, disclose system limitations and intended uses, provide documentation on performance and risk mitigation, and ensure synthetic content (e.g., deepfakes) is identifiable—unless exempt (e.g., artistic works, text for public interest).
- Requires deployers to implement a risk management policy, complete impact assessments before and after deployment, disclose AI use to consumers in plain language, and provide explanations for adverse decisions—including how the AI contributed and what data it used.
- Grants consumers a private right of action to sue for violations, with potential remedies including injunctions and attorney fees; developers or deployers can avoid liability if they cure violations within 45 days of discovery.
- Exempts certain uses and sectors—including federal government systems, insurance (if regulated), financial institutions under existing oversight, research sandboxes, and health care providers acting under HIPAA—while maintaining core prohibitions on algorithmic discrimination.
Who is affected
- AI developers — Businesses that develop or significantly modify high-risk AI systems and earn over $100,000 in annual Washington revenue must provide detailed disclosures, documentation, and support for impact assessments, and may face civil liability for algorithmic discrimination.
- AI deployers — Businesses that use high-risk AI systems to make consequential decisions (e.g., hiring, housing, credit, education, legal outcomes) must implement risk management policies, conduct impact assessments, disclose AI use to consumers, and provide explanations for adverse decisions.
- Consumers — Washington residents making personal, household, or consumer-facing decisions (e.g., applying for jobs, loans, housing, or education) gain new rights to know when AI is used, understand how it affects them, and receive explanations for adverse outcomes.
- Insurers and financial institutions — Insurance companies and financial institutions may be exempt if they are already regulated under existing state or federal rules that address algorithmic discrimination in insurance or financial services.
Pro/Con Analysis
Stronger case for benefits
Potential Benefits (5)
Consumers gain a statutory right to receive clear, plain-language explanations for adverse consequential decisions made by AI—including how the system contributed and what data it used—empowering individuals to challenge errors or bias in housing, employment, credit, and legal outcomes.
Rights & LibertiesPeopleRef: Sec. 3(5)Developers must disclose system limitations, known risks of algorithmic discrimination, and mitigation steps—enabling consumers and downstream deployers to make informed choices and hold developers accountable for biased or opaque systems.
Rights & LibertiesPeopleRef: Sec. 2(2)(a)-(c)Deployers must disclose in plain language when AI is used to make consequential decisions—including what personal attributes are assessed and how—giving Washingtonians meaningful notice and agency in digital interactions.
Rights & LibertiesPeopleRef: Sec. 3(4)Both developers and deployers owe a duty of reasonable care to prevent algorithmic discrimination, with a rebuttable presumption of compliance for those meeting documentation and risk-mitigation requirements—creating strong legal leverage against biased systems in hiring, housing, and lending.
Rights & LibertiesPeopleRef: Sec. 2(1), Sec. 3(1)Requirements to make synthetic content (e.g., deepfakes) identifiable and to explain AI’s role in adverse decisions reduce deception risks in critical domains like legal proceedings, immigration, and public benefits—protecting vulnerable populations from automated misclassification.
Public SafetyPeopleRef: Sec. 2(7)(a)-(c), Sec. 3(5)
Potential Concerns (5)
Businesses must disclose AI use and provide detailed explanations for adverse decisions (e.g., hiring, credit, housing denials), increasing operational complexity and legal exposure—especially for small firms without legal or compliance staff.
Business & EmploymentPeopleRef: Sec. 3(4)-(5)Deployers must complete and maintain impact assessments before deployment and after significant updates—including raw data retention for three years post-deployment—creating recurring compliance costs that disproportionately burden small and mid-sized businesses without dedicated AI governance teams.
Business & EmploymentPeopleRef: Sec. 3(3)(a)-(ix)Generative AI developers must ensure synthetic content is identifiable using tools accessible to consumers, which may require embedding detection watermarks or metadata—costly for small developers and potentially incompatible with existing creative workflows or third-party platforms.
Business & EmploymentLean peopleRef: Sec. 2(7)(a)-(c)The private right of action allows consumers to sue for violations, but the 45-day cure window and rebuttable presumption of reasonable care for compliant actors may favor plaintiffs with resources to identify and exploit technical noncompliance, while exposing businesses to litigation risk even when acting in good faith.
Rights & LibertiesLean peopleRef: Sec. 5While the bill allows reliance on existing regulatory impact assessments (e.g., from insurance or financial regulators), alignment with those frameworks is not guaranteed—creating uncertainty and potential duplication for firms operating across multiple regulated sectors.
Business & EmploymentRef: Sec. 3(3)(c)(ii)
Who Is Most Affected
Low- and middle-income consumers applying for jobs, housing, credit, or public benefits gain enforceable rights to transparency and redress if AI systems produce biased or erroneous outcomes—especially important for communities historically subject to discriminatory algorithms.
Small and mid-sized businesses deploying AI in hiring, lending, or customer service face new documentation, assessment, and disclosure obligations—though they benefit from clear safe harbors and the cure window; larger firms with compliance infrastructure may adapt more easily.
AI developers (especially startups and independent tool builders) must invest in documentation, risk assessments, and synthetic-content identification—potentially raising barriers to entry, though the $100K revenue threshold protects micro-businesses.
Insurers and financial institutions already subject to state or federal oversight (e.g., by the Office of the Insurance Commissioner or OCC) are largely exempt—preserving existing regulatory frameworks and reducing compliance burden for these sectors.
Tech platforms and large employers using AI at scale will face significant operational changes—including impact assessments, consumer disclosures, and potential litigation—but have the resources to comply; they may pass some costs to consumers through higher fees or reduced service options.